Dell P. Dunn MD, Presenter: Nothing to Disclose

Olga Rachel Brook MD, Abstract Co-Author: Research Grant, Guerbet SA

Alexander Brook PhD, Abstract Co-Author: Spouse, Research Grant, Guerbet SA

Giselle Y. Revah MD, Abstract Co-Author: Nothing to Disclose

Sumayya Shamaroze Jawadi MD, Abstract Co-Author: Nothing to Disclose

Karen Sisi Lee MD, Abstract Co-Author: Nothing to Disclose

Maryellen Ruth Morris Sun MD, Abstract Co-Author: Investigator, Bracco Group

Koenraad J. Mortele MD, Abstract Co-Author: Nothing to Disclose

The purpose of this study is to assess inter-observer variability in measurements of pancreatic cystic lesions on MRI and its impact on patient management using current guidelines.

In this IRB-approved, HIPAA-compliant study, 144 MRI examinations in patients with pancreatic cystic lesions were reviewed by 2 fellowship-trained abdominal imagers, each with 6 yrs of experience, and 2 abdominal imaging fellows. The studies were selected to provide an even distribution of pancreatic cystic lesions ranging in size from 5 to 35mm. The reviewers were instructed to record the single largest measurement according to their standard clinical practice using the plane and sequence of their choice. These measurements were then placed into 4 categories delineated by current management guidelines.

There was moderate overall agreement (kappa = 0.55) for the four readers, however all 4 readers placed the cystic lesion into the same management category in only 65/145 cases (45%). Pairwise comparison of agreement between individual readers yielded kappa ranging from 0.46 to 0.61, with the highest agreement found between the two most experienced readers and the least agreement between the two less experienced readers. The intraclass correlation coefficient, a measure of quantitative agreement on lesion size, showed strong agreement with a value of 0.80. The within-subject standard deviation increased with increasing lesion size, ranging from 2.7mm±0.3mm for lesions <1cm up to 5.9mm±0.5mm for lesions over 3cm.

There is significant variability in the measurement of cystic pancreatic lesions on MRI that could lead to erroneous reporting of growth or unwarranted changes in management.

MRI measurements of pancreatic cysts can vary significantly between readers with impact on patient management. Standardization of measurement technique may reduce variability and should be explored.

Dunn, D, Brook, O, Brook, A, Revah, G, Jawadi, S, Lee, K, Sun, M, Mortele, K, The Importance of Measuring Pancreatic Cystic Lesions Accurately on MRI: An Assessment of Inter-observer Variability and Impact on Patient Management.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13044250.html